mx-toolbox

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's workflow (SKILL.md) instructs the agent to run Membrane commands like membrane action list and membrane request ... against the Mx Toolbox API (https://mxtoolbox.com/api/), which fetches external third-party API responses that the agent must read and use (e.g., action IDs/input schemas) to decide and invoke subsequent actions, exposing it to untrusted third-party content.